Technical Field
The present invention relates generally to information processing and, in particular, to shadow agent projection in multiple places to reduce agent movement over nodes in a distributed agent-based simulation.
Description of the Related Art
Recently, agent-based social simulations have been utilized to support the decision making of city planners for various social issues including evacuation and intelligent.
To evaluate complex behavior with interactions among heterogeneous people in a large city, such a social simulation needs to manage a large number (e.g., millions) of agents with various behavior models and preferences. Moreover, computation speed is also required to repeatedly analyze enormous combinations of possible situations and strategies.
Thus, the distributed parallel execution of large-scale agent-based social simulations is very important for evaluating the complex social interaction of an enormous number of people and situations with acceptable performance.
X10-based Agent Simulation on Distributed Infrastructure (XASDI) developed by IBM® is a platform to realize the distributed parallel execution of large-scale social simulation with X10.
X10 denotes a distributed node as a “Place”, and provides functionalities of a global address space, parallel execution, and synchronization among multi-Places. By using X10, XASDI manages distributed agents and realizes the message and agent transfer between Places.
Realistic social agents tend to include a large amount of data that, in turn, includes demographics, preferences, and history. Hence, the transfer of such an agent incurs a heavy communication cost that has an adverse effect on performance.
Accordingly, there is a need for an improved approach to transfer an agent.